The Fantasy Premier League (FPL) is a very popular game, our job, as players, is to build a team with Premier League teams players, filling all position, goalkeepers, defense, midfielders and forwards, and even the bench, all with a starting budget of £100, other rules apply as, no more of 3 players from the same team, and we need a specific number of player of each position.
Previous to each matchweek, we may make a transfer, if we want to make more than one we will suffer a points penalty. We also need to select the formation and which player will play and which will be benched, also the captain (who gets double points).
Then, with each game the player gets points, based on what they accomplished, like goals, assists, clean sheets, etc. Finally we sum all points our team makes. It’s an easy game, but is really fun and addictive, and of course, there are a lot of things we can use to optimize our score using data.
Recently I discovered the official API of the game, with a lot of very interesting data in real time.
How does the international break effect on players performance? Does players with international duty improve their performance after the international break?
To improve my transfer suggestion system I train a ML model to predict each player score on next gameweek.
A brief look into the season performance of all players and teams
It's time to make data based transfers to improve the team value of the season.
Is the dreamteam the best team every gameweek? Now I will compare the gameweek dreamteam with the FPL dreamteam
I started to ask myself how to improve my game decisions, then discovered the FPL API while investigating , now lets see what I can do.